Adoption and Implementation of Virtualization Versus Physical Server Infrastructure

The Basics of Virtualization

Virtualization is the ability to run one or more virtual machines on top of one physical host (Ruest & Ruest, 2009).  With virtualization, it is easy to spin up testing, training, and development environments without the need for separate hardware.  Virtualization has been a disruptive force within the industry, and it has significantly impacted the way data centers operate.  In data centers, virtualization introduced vast efficiency and cost improvements.  Before virtualization, it was common for data centers to run many dedicated servers at minimal utilization rates.  Utilization rates were low because the hardware had to handle potential surges in usage that only occasionally occurred.  Enough capacity also had to be provisioned to account for projected growth.  After introducing virtualization, the hardware efficiently increased because customers could use Virtual Private Servers (VPSs) rather than a dedicated server.  Multiple VPSs could run on the hardware that previously was dedicated for a single purpose.  These VPSs provide the same reliability and security as a dedicated server.  Virtualization is not limited to data centers (server virtualization).  Virtualization has different forms, including desktop virtualization, in which multiple virtual desktop operating systems can run on a single machine.  Other types of virtualization include application virtualization, storage virtualization, and network virtualization.  

The Technical Specifications of Virtual Machines

System requirements for virtual machines (VMs) vary depending on the purpose, but the architecture of all virtual machines requires a hypervisor.  A hypervisor is software that creates and runs virtual machines (VMs) (Redhat, 2020).  The hypervisor software isolates the operating systems and resources for the VMs so that multiple operating systems can share the same physical computing resources (IBM, n.d.).  This software also provides management functionality for the VM and allocates the necessary system resources to each VM.  These system resources include memory, drivers, process schedulers, and other requirements that each VM needs to function.  There are two categories of hypervisors: Type 1 hypervisors and Type 2 hypervisors.  Type 1 hypervisors run directly on the physical hardware and do not require an operating system (OS).  Type 2 hypervisors, on the other hand, run as an application on an OS (IBM). 

Virtual machines contain all of the devices and capabilities that are in a physical system.  This includes storage, BIOS, memory, ports, processor, networking capabilities, etc.  The VM configuration for common configurations can be chosen from a menu of pre-configured options.  This allows for configuration and spin-up of the needed VM can using just a few clicks.  The user also can build the VM technical specification from scratch by choosing custom options or modifying the options of a pre-configured VM. 

Physical Data Center Cost Considerations

The capital expenditure required to purchase, implement, operate, and maintain a physical data center is significant.  The data center’s size depends on the organization’s requirements and could range from a small 500 square foot room to over 5000 square feet for a mid-sized organization.  Beyond size, the required uptime (availability) for the data center impacts the cost significantly.  This requirement affects the design, construction, and components required for the data center.  If the organization can withstand system outages, there would be cost savings based on the reduced uptime requirement.

In many cases, a high level of system availability is critical for the organization.  This criticality requires the implementation of redundancy in the form of Uninterruptible power supplies (UPS) and backup power (generators).  Network redundancy may also be necessary to ensure that the system does not suffer an outage in the case of a fiber cut or network failure.  High levels of network redundancy are achieved by constructing two distinct communications paths to the datacenter.  In the case of a network failure, the system can automatically switch to the redundant network path without significantly impacting the users.  The estimated cost expenditures required to construct a physical data center ranges from tens of millions to hundreds of millions of dollars, depending on the size and requirements.

Virtualization Cost Considerations

As stated, a significant advantage of virtualization is cost savings.  The most obvious cost savings is through the reduction of underutilized physical hardware components.  The cost savings from the decrease in the need for physical hardware is amplified by less need for climate-controlled space, redundancy capabilities, and power.  Simply put, reduction in hardware has significant cost savings implications in multiple areas. 

Virtualization technology enables the business model for cloud computing providers.  By definition, cloud computing is the on-demand delivery of IT resources over the Internet with pay-as-you-go pricing (AWS).  Cloud computing relies on providers rely on virtualization and enables organizations to purchase IT services as an operating expense than investing in the costs associated with a physical data center.  The cloud model shifts the capital expenditures from the user to the cloud provider.  This business model works because the cloud providers utilize virtualization to serve many separate customers with the same hardware.  This allows the cloud provider to realize economies of scale to justify the high costs of deploying physical data centers.

Adoption of Virtualization

Virtualization technology began to become widely adopted in the early 2000s (Red Hat, 2018).  Since then, the adoption of virtualization has steadily increased.  Virtualization adoption was rapidly changed with the development of the cloud.  In effect, the cloud was born when virtualization meets the Internet (IBM, 2017).  Amazon Web Services (AWS) launched the first public cloud in 2002.  Since 2012, the cloud has evolved and added many cloud hardware options, operating systems, and software options.  In the first quarter of 2021, AWS reported $13.5 billion in sales.  This was an annual revenue increase of 32%.  These statistics show that virtualization adoption (in the cloud computing sector) is exceptionally rapid.  Large enterprises and governmental organizations are adopting cloud technologies to replace aging data centers so that they can realize the cost savings and take advantage of the agility offered by virtualized cloud environments.  Smaller organizations such as single-office small businesses also adopt the cloud to avoid the capital expenditures required for on-premises equipment.

Beyond the cloud, desktop virtualization is also widely adopted.  Companies such as VMware and Oracle offer virtualization software that allows security researchers, software developers, and others to create VMs.  VMs are commonly used for testing new operating systems, investigating malware, running incompatible software, or for secure internet browsing (IBM Cloud Education, 2019)

References

A brief history of cloud computing BM Cloud Team, (2017, January 6).  https://www.ibm.com/cloud/blog/cloud-computing-history

Hypervisers. (n.d.). IBM. Retrieved October 34, 2021 from https://www.ibm.com/cloud/learn/hypervisors

Redhat (2018) What is virtualization? https://www.redhat.com/en/topics/virtualization/what-is-virtualization

Redhat (2020, January 10) What is a hypervisor? https://www.redhat.com/en/topics/virtualization/what-is-a-hypervisor

Ruest, N. & Ruest, D., (2009) Virtualization, a beginner’s guide. McGraw-Hill Education

Photo of author

Donald Korinchak, MBA, PMP, CISSP, CASP, ITILv3

Donald Korinchak is a Cybersecurity Professional in the Washington DC area. Donald holds an MBA from the University of Pittsburgh Katz School of Business. Donald is considered a thought leader in business, leadership, and cybersecurity issues.

Leave a Comment